🚢 Titanic Survival Prediction using Machine Learning Pipeline

This project demonstrates the use of Machine Learning Pipelines in Python using scikit-learn to predict survival on the Titanic dataset. The pipeline handles all preprocessing steps and applies a classifier in a clean, reproducible way.

---

📁 Project Structure

├── Machine pipeline.ipynb             # Main notebook with model building pipeline
├── predict using pipeline.ipynb       # Notebook to use trained pipeline for predictions
├── pipe.pkl                           # Trained pipeline saved as a pickle file
├── README.md                          # Project documentation

---

🚀 Features

• Complete ML pipeline including preprocessing and model training
• Handling missing values and categorical encoding
• Pipeline serialization using joblib
• Inference using the saved pipeline
• Simple and extendable structure

---

📦 Requirements

Install dependencies using:

pip install -r requirements.txt

You’ll need:

• scikit-learn
• pandas
• numpy
• joblib
• matplotlib (optional for visualizations)

---

📊 Dataset

The dataset used is the classic Titanic dataset.
It includes features such as Pclass, Sex, Age, Fare, and survival labels (Survived).

---

🧠 Model Pipeline

The pipeline includes the following steps:

1. Imputation: Filling missing values (e.g., age, embarked).
2. Encoding: Converting categorical variables (Sex, Embarked) using OneHotEncoding.
3. Feature Scaling: StandardScaler for numeric features.
4. Feature Selection: (Optional) using SelectKBest.
5. Classification: Using RandomForestClassifier.

---

🛠 How to Use

1. Train the model: Open Machine pipeline.ipynb and run all cells. This notebook creates the pipeline, trains it, and saves it to pipe.pkl.

2. Predict using the saved model: Open predict using pipeline.ipynb to load the trained model and make predictions on new or test data.

---

🔍 Example Prediction

import joblib
import pandas as pd

pipe = joblib.load("pipe.pkl")
new_data = pd.DataFrame([{
    "Pclass": 3,
    "Sex": "male",
    "Age": 22,
    "Parch": 0,
    "Embarked": "S"
}])
prediction = pipe.predict(new_data)
print("Survived" if prediction[0] == 1 else "Did not survive")

---

📚 Learn More

• scikit-learn Pipelines
• Kaggle Titanic Competition

---

🙌 Acknowledgements

• Kaggle for the dataset.
• scikit-learn for the pipeline and modeling tools.